NBER WORKING PAPER SERIES BANKING CRISES: AN EQUAL OPPORTUNITY MENACE Carmen M. Reinhart Kenneth S. Rogoff Working Paper 14587 http://www.nber.org/papers/w14587 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 December 2008 The authors are grateful to Vincent Reinhart, Keyu Jin, Tarek Hassan, Vania Stavrakeva for useful comments and suggestions on an earlier draft, and to Cesar Sosa, Chenzi Xu and Jan Zilinsky for excellent research assistance. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer- reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications. © 2008 by Carmen M. Reinhart and Kenneth S. Rogoff. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. Banking Crises: An Equal Opportunity Menace Carmen M. Reinhart and Kenneth S. Rogoff NBER Working Paper No. 14587 December 2008 JEL No. E6,F3,N0 ABSTRACT The historical frequency of banking crises is quite similar in high- and middle-to-low-income countries, with quantitative and qualitative parallels in both the run-ups and the aftermath. We establish these regularities using a unique dataset spanning from Denmark's financial panic during the Napoleonic War to the ongoing global financial crisis sparked by subprime mortgage defaults in the United States. Banking crises dramatically weaken fiscal positions in both groups, with government revenues invariably contracting, and fiscal expenditures often expanding sharply. Three years after a financial crisis central government debt increases, on average, by about 86 percent. Thus the fiscal burden of banking crisis extends far beyond the commonly cited cost of the bailouts. Our new dataset includes housing price data for emerging markets; these allow us to show that the real estate price cycles around banking crises are similar in duration and amplitude to those in advanced economies, with the busts averaging four to six years. Corroborating earlier work, we find that systemic banking crises are typically preceded by asset price bubbles, large capital inflows and credit booms, in rich and poor countries alike. Carmen M. Reinhart University of Maryland School of Public Policy and Department of Economics 4105 Van Munching Hall College Park, MD 20742 and NBER [email protected] Kenneth S. Rogoff Thomas D Cabot Professor of Public Policy Economics Department Harvard University Littauer Center 232 Cambridge, MA 02138-3001 and NBER [email protected] I. Introduction Until very recently, the study of banking crises has typically focused either on earlier historical experiences in advanced countries, mainly the banking panics before World War II, or else has focused on modern-day emerging market experiences.1 This dichotomy is perhaps shaped by the belief that for advanced economies, destabilizing, systemic, multi-country financial crises were a relic of the past.2 Of course, the recent global financial crisis emanating out of the United States and Europe has dashed this misconception, albeit at great social cost. As this paper will demonstrate, banking crises have long been an equal opportunity menace. We develop this finding using a core sample of sixty-six countries (plus a broader extended sample for some exercises). 3 We examine banking crises ranging from Denmark’s financial panic during the Napoleonic War to the current “first global financial crisis of the 21st century.” The incidence of banking crises proves to be remarkably similar in the high- and middle-to-low-income countries. Indeed, the tally of crises is particularly high for the world’s financial centers: the United Kingdom, the United States, and France. Perhaps more surprising still are the qualitative and quantitative parallels across disparate income groups. These parallels arise despite the relatively pristine modern sovereign default records of the rich countries. Three features of our expansive dataset are of particular note. First, our data on global banking crises go back to 1800, extending the careful study of Bordo, et al. (2001) that covers 1 See Calomiris and Gorton (1991) and Gorton (1988) on pre–WWII banking panics; Sundararajan and Baliño (1991) for several emerging market case studies; Jácome (2008) on banking crises in Latin America. 2 Studies that encompass episodes in both advanced and emerging economies include Bordo et al. (2001), Demirgüç-Kunt and Detragiache (1998) and Kaminsky and Reinhart (1999). 3 The core sample spans 66 advanced and emerging market economies in Africa, Asia, Europe, Latin and North America and Oceania; see Appendix Table A1. The extended sample includes all countries, see Table A3. 1 back to 1880. Second, to our knowledge, we are the first to examine the patterns of housing prices around major banking crises in emerging markets, including Asia, Europe and Latin America. Our emerging market data set facilitates comparisons, across both duration and magnitude, with the better-documented housing price cycles in the advanced economies, which have long been known to play a central role in financial crises.4 We find that real estate price cycles around banking crises are similar in duration and amplitude across the two groups of countries. This result is surprising given that almost all other macroeconomic and financial time series (income, consumption, government spending, interest rates, etc.) exhibit higher volatility in emerging markets.5 Third, our analysis employs the comprehensive historical data on central government tax revenues and debt compiled in Reinhart and Rogoff (2008a,c). These new data afford a new perspective on the tax and debt consequences of the banking crises (Previously, the kind of historical data on debt necessary to analyze the aftermath of banking crises across countries was virtually non-existent for years prior to 1990.6) We find that banking crises almost invariably lead to sharp declines in tax revenues as well significant increases in government spending (a share of which is presumably dissipative). On average, government debt rises by 86 percent during the three years following a banking crisis. These indirect fiscal consequences are thus an order of magnitude larger than the usual bank bailout costs that are the centerpiece of most previous studies. That fact that the magnitudes are comparable in advanced and emerging market economies is also quite remarkable. Obviously, both the bailout costs and the fiscal costs depend on a host of political 4 See Reinhart and Rogoff (2008b) for an analysis of all post–WWII banking crises in advanced economies. 5 See, for instance, Agénor, McDermott, and Prasad (2000). 6 Bordo and Meissner (2006) offer domestic debt data for selected years across 30 countries for 1880–1913, while Jeanne and Guscina (2006) provide domestic debt for 19 countries for 1980–2005. The Reinhart and Rogoff (2008c) time series for sixty-six countries spans 1913–2007, and much earlier for a large subset of these countries. 2 and economic factors, including especially the policy response as well as the severity of the real shock which, typically, triggers the crisis.7 The paper proceeds as follows. Section II provides an overview of the history of banking crises, with particular emphasis on the post-1900 experience. We also document the incidence and frequency of banking crises by country and by region. We discuss the links between banking crises, financial liberalization, the degree of capital mobility, and sovereign debt crises and discuss international financial contagion. Section III examines some of the common features in the run-up to banking crises across countries and regions over time. The focus is on the systematic links between cycles in international capital flows, credit, and asset prices—specifically, home and equity prices. The next section examines some of the common features of the aftermath of banking crises. We document the toll that the crisis takes on output and government revenues, as well as the typically profound effect on the evolution of government debt during the years following the crisis. The concluding section takes up the question of “graduation.” Specifically, to what extent do countries ever “graduate” from (stop experiencing) serial major financial crises as they seem to graduate from serial sovereign debt crises?8 7 Reinhart and Rogoff 2008a,c show that output growth typically decelerates in advance of a crisis. 8 An example of graduate from serial default is France, which defaulted 8 times on its external debt between 1500 and 1800, but has not defaulted since. 3 II. Banking Crises in Historical Perspective We begin this section by providing an overview of the evolution of banking crises through history. To do so, it is necessary to first identify and date banking crisis episodes. Our approach, which follows the standard methodology in the literature (e.g., Kaminsky and Reinhart, 1999, Bordo, et al., 2001, and Caprio and Klingebiel, 2005, among others), is documented in detail in the appendix, along with our principal bibliographical sources. 9 One dimension that distinguishes this study from previous efforts is that our dating of crises extends far before the much-studied modern post–World War II era. Specifically, we start in 1800. Our work was greatly simplified back to 1880 by the careful study of Bordo, et al. (2001), but for the earlier period we had to resort to old and often obscure works. The earliest advanced-economy banking crisis in our sample is France 1802; early crises in emerging markets befell India, 1863, China (several episodes during the 1860s–1870s), and Peru in 1873. 10 It may come as a surprise to the reader that previous attempts to document banking crises for the pre–World War II period are so limited.